The Role:

You’ll help turn Applied AI ideas into dependable software that gets used in the real world — by customers and by our internal threat analysts. Your work will help the team move faster (better tooling and workflows), keep results consistent and reproducible (evaluation and tracking), and help analysts self-serve through well-designed prompts, context, and tools. 

You’ll be supported by senior engineers, but you’ll still ship meaningful work and see direct impact. 

What you’ll do:

• Build practical Python software 
  Create and improve backend components, libraries, and CLI tools used across our Applied AI Research workflows. 

• Help analysts self-serve LLM workflows 
  Partner with internal threat analysts to develop and maintain prompt/context patterns, guardrails, and templates — and help roll these into a self-serve setup using tools like Langfuse (or similar) for prompt management, tracing, and evaluation. 

• Contribute to productizing and deploying AI solutions 
  Work with senior engineers to move prototypes into production: packaging, integration, reliability, and deployment. You’ll take part in discussions about tradeoffs (speed vs. quality, cost vs. latency, build vs. buy) and learn how production decisions get made. 

• Support evaluation & reproducibility 
  Contribute to lightweight evaluation harnesses, experiment tracking, and documentation so we can repeat results and trust what we ship. 

• Use AI coding assistants thoughtfully 
  Use tools like Claude Code, Windsurf, Cursor (or similar) to accelerate development while keeping a high bar for correctness, readability, and tests. 

What you bring:

(We know no one checks every box — this is our ideal list.) 

• 1–2 years of experience in software development (or equivalent through internships + strong projects). 

• Good software fundamentals: clean code, debugging, data structures, basic design/architecture thinking. 

• Python, backend-leaning: comfortable writing small services, scripts, or CLI apps; familiarity with APIs and data handling. 

• Interest in applied AI: exposure to AI tooling/data pipelines/context building (coursework, projects, or on-the-job). 

• Hands-on with AI coding assistants: experience using at least one of Claude Code, Windsurf, Cursor, or similar. 

• Proof you like building: side projects, GitHub repos, demos, or hobby tools you can share. 

Nice to have:

• Interest in agentic AI (tool use, workflows, simple orchestration). 

• Familiarity with Docker, basic cloud concepts, or evaluation/observability tools. 

• Comfort collaborating with non-engineers (e.g., analysts) and translating needs into tooling. 

How we’ll track progress:

In your first 3–12 months, we’ll look for: 

Shipped improvements to our internal AI tooling (e.g., a CLI tool, pipeline step, evaluation feature, or reliability enhancement). 

Analyst enablement: prompt/context patterns and a workflow that makes analysts faster and more consistent (with tracing/eval via Langfuse or similar). 

Production contribution: you help deliver at least one AI feature/workflow into a production-like environment with guidance (and you can explain the tradeoffs we chose). 

Growing ownership: you get faster at scoping work, asking good questions, and delivering maintainable code.